Doc.: IEEE 802.11-05/0946r0 Submission September 2005 Douglas S. Chan et al., Cornell UniversitySlide 1 Improving IEEE 802.11 Performance with Cross-Layer.

September 2005

Douglas S. Chan et al., Cornell University

Slide 1

doc.: IEEE 802.11-05/0946r0

Submission

Improving IEEE 802.11 Performance with Cross-Layer Design and Multipacket Reception via Multiuser Iterative Decoding

Notice: This document has been prepared to assist IEEE 802.11. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein.

Release: The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contribution may be made public by IEEE 802.11.

Patent Policy and Procedures: The contributor is familiar with the IEEE 802 Patent Policy and Procedures <http:// ieee802.org/guides/bylaws/sb-bylaws.pdf>, including the statement "IEEE standards may include the known use of patent(s), including patent applications, provided the IEEE receives assurance from the patent holder or applicant with respect to patents essential for compliance with both mandatory and optional portions of the standard." Early disclosure to the Working Group of patent information that might be relevant to the standard is essential to reduce the possibility for delays in the development process and increase the likelihood that the draft publication will be approved for publication. Please notify the Chair <[email protected]> as early as possible, in written or electronic form, if patented technology (or technology under patent application) might be incorporated into a draft standard being developed within the IEEE 802.11 Working Group. If you have questions, contact the IEEE Patent Committee Administrator at <[email protected]>.

Date: 2005-09-20

Name Company Address Phone email Douglas S. Chan [email protected]

Prapun Suksompong

[email protected]

Jun Chen

607-254-8818

[email protected]

Toby Berger

Cornell University

School of Electrical and Computer Engineering, Cornell University, Ithaca, NY 14853

607-255-1447 [email protected]

Authors:

September 2005

Douglas S. Chan et al., Cornell University

Slide 2

doc.: IEEE 802.11-05/0946r0

Submission

Abstract

Receivers today have the ability to decode more than one packets from multiple users. Such a physical layer can deliver significant improvements to network performances. Thus, the classical collision model is no longer realistic and a cross-layer approach should be employed when designing multiple access protocols. This is especially the case for CSMA communications, which previously have not been implemented with a multipacket reception (MPR) model. We propose applying recent information theoretic results in multiuser iterative decoding to help improve IEEE 802.11 wireless LAN standards’ performances. Our method also preserves the underlying physical layer’s implementation.

September 2005

Douglas S. Chan et al., Cornell University

Slide 3

doc.: IEEE 802.11-05/0946r0

Submission

Collision Channel Model

• Traditionally, network research done over collision channel model – Only one transmission can occur over channel at any instance

• Analysis is more tractable, reflects PHY layers at the time, worst case scenario

BackoffCollision

STA B

Channel activities

STA A

Success

September 2005

Douglas S. Chan et al., Cornell University

Slide 4

doc.: IEEE 802.11-05/0946r0

Submission

Multipacket Reception Channel Model

• Receivers today can correctly separate and decode multiple packets transmitted simultaneously– e.g., by signal processing or spread spectrum modulation

• This is referred to as a multiple packet reception (MPR) channel [Ghez]

• Better performance than collision channel and more realistic

BackoffTwo successes

STA B

Channel activities

STA A

Success

September 2005

Douglas S. Chan et al., Cornell University

Slide 5

doc.: IEEE 802.11-05/0946r0

Submission

CSMA with Multipacket Reception • CSMA over an MPR-capable PHY layer has not been

previously studied

• First proposed recently by Chan, Berger and Tong [Chan1] in 2004– Showed that MPR improves performance of CSMA over Collision

Channel

– Performance also improves over S-ALOHA

– However, as MPR strength becomes stronger, carrier sensing and scheduling becomes unnecessary• But requires a lot of resource for MPR strength to be perfect

• So this is a balance between MPR-PHY strength and scheduling

September 2005

Douglas S. Chan et al., Cornell University

Slide 6

doc.: IEEE 802.11-05/0946r0

Submission

CSMA Performance Improvement with MPR

September 2005

Douglas S. Chan et al., Cornell University

Slide 7

doc.: IEEE 802.11-05/0946r0

Submission

Current PHY layers for 802.11• Currently, their PHY implementations actually can support multiuser

detection– 802.11: frequency hopping, direct sequence spread spectrum– 802.11b: DSSS with higher rate codes (and optional PBCC)– 802.11a: OFDM– 802.11g: OFDM (and optional PBCC)– 802.11n(WG): MIMO

• Used these MPR implementations to– 1. Combat narrowband interference in the unlicensed radio-bands

where these standards and other RF applications co-exist– 2. Mitigate interferences between nearby WLANs in case their chosen

channels’ bandwidths partially or wholly overlap

• The MPR abilities are used to separate transmissions in unrelated nearby WLANs as opposed to increasing the multiaccess transmission capacity of a single WLAN.

September 2005

Douglas S. Chan et al., Cornell University

Slide 8

doc.: IEEE 802.11-05/0946r0

Submission

CSMA Performance Improvement with MPR

September 2005

Douglas S. Chan et al., Cornell University

Slide 9

doc.: IEEE 802.11-05/0946r0

Submission

Implementing MPR on 802.11• MPR can improve performance of 802.11

• Consider adding MPR capability to next generations of 802.11

• But MPR usually involves modifications of PHY layer– Makes it harder to maintain backwards compatibility

– Or ready application to current PHY proposals

• Ideally we would like an MPR solution that can preserve the underlying PHY layer

September 2005

Douglas S. Chan et al., Cornell University

Slide 10

doc.: IEEE 802.11-05/0946r0

Submission

Survey of methods for MPR• Direct Sequence Spread Spectrum with (pseudo-)orthogonal codes

– PHY centric/Requires changes in PHY

• MIMO – Space-Time Block Coding– Embedding known symbols in packets and successive cancellation

– PHY centric/Requires changes in PHY

• Successive Cancellation– 1. Signaling level

• (Pseudo-) Orthogonal waveforms– PHY centric/Requires changes in PHY

– 2. Modulation/Coding level• Treat other users’ signals as noise and successively decode

– Coding centric/May not require changes in PHY

• Joint Decoding with Multiaccess codes:– Iterative detection and decoding– Iterative joint decoding

– Coding centric/May not require changes in PHY

September 2005

Douglas S. Chan et al., Cornell University

Slide 11

doc.: IEEE 802.11-05/0946r0

Submission

Network model for multiaccess codes• User i encodes source data via his/her assigned codebook as Xi

– This codebook is part of an overall codebook for L users

– Central receiver has overall codebook

• Central receiver receives channel affected version of symbols sent by multiple users:

– Y is received signal

– Xi is symbol transmitted by user i

– Hi is channel coefficients for user I

– N is noise

• Central receiver decodes the data sent by each user via:– Joint decoding or successive cancellation

• Scheme works as long as signaling scheme over channel is additive

September 2005

Douglas S. Chan et al., Cornell University

Slide 12

doc.: IEEE 802.11-05/0946r0

Submission

Signaling for multiaccess codes• Examples of additive signaling schemes

– Amplitude modulation

– Phase modulation: BPSK, QPSK, n-QAM (!!)

• Not additive: If information encoded in differential of phase (eg. DPSK)• 802.11 PHYs all employ non-differential phase modulation, i.e. additive

1 1 -1 1

-1 1 1 1

-1 -1 1 1

-1 1 1 3

X1

X2

X3

Channel

Noise

Hard/SoftDecision

MUD

Receiver

X1= 1 1 -1 1

X2= -1 1 1 1

X3= -1 -1 1 1

September 2005

Douglas S. Chan et al., Cornell University

Slide 13

doc.: IEEE 802.11-05/0946r0

Submission

New MUD layer• Propose adding new “Multiuser

Decoding” (MUD) layer to perform multiuser detection and decoding

• Can employ any underlying PHYs to achieve multiaccess coding

• MAC still needed for scheduling re-transmissions

• Can turn on and off as wished

• Joint decoding performs at least as good as successive decoding– Recent results on iterative joint

decoding show very good performance

– Can focus implementation on iterative joint decoding codes

MUD

PHY

MAC

Performs Multiuser detection and decoding

Current 802.11Proposed 802.11

September 2005

Douglas S. Chan et al., Cornell University

Slide 14

doc.: IEEE 802.11-05/0946r0

Submission

MUD 1: Graph-based jointly decoded codes• Introduced in Palanki, Khandekar and McEliece’s 2001 paper [Palanki]

• Each user encodes with their own distinct low density parity check LDPC codebook

• Use belief propagation (BP) to perform iterative joint decoding – Each edge is an independent estimate of each Xi

– Contains joint channel information

– Iteration stops until all parity check satisfied or after certain iterations

Variablenodes

Variablenodes

Checknodes

September 2005

Douglas S. Chan et al., Cornell University

Slide 15

doc.: IEEE 802.11-05/0946r0

Submission

MUD 1: Graph-based jointly decoded codes• Design of multiuser LDPC codes can be achieved with a graph splitting and density evolution

techniques

• Advantages:– Each user’s code can be of its distinct rate

• Achieving priority scheduling (QoS)

– BP decoding of LDPC can be done in parallel

• Implementation issues:– Requires fixed length codewords– Codes designed only for a fixed number of users

• We can put limit to the number of users associating to AP

– AP needs to tell each user which codebook to use during association– Requires good codes to combat code expansion (CE)– No MPR when AP transmits

September 2005

Douglas S. Chan et al., Cornell University

Slide 16

doc.: IEEE 802.11-05/0946r0

Submission

MUD 2: Iterative detection and decoding• Sanderovich, Peleg and Shamai’s 2005 paper [Sanderovich]

• Each user encodes with same LDPC code

• Performs user-specific random interleaving to codeword

10

00

Inte

rlea

ver

Inte

rlea

ver

LDPC

LDPC

QPSK 1 1 -1 1

-1 -1 1 1

-1 -1 1 1

Inte

rlea

ver

11

LDPC

September 2005

Douglas S. Chan et al., Cornell University

Slide 17

doc.: IEEE 802.11-05/0946r0

Submission

MUD 2: Iterative detection and decoding• Then performs iterative but separate multiuser detection and single-user

decoding– Multiuser detection can be an optimal soft-decision LLMSE

– Single user detection is single LDPC BP decoder

– Iteration stops until all parity check satisfied or after certain iterations

Mul

tius

er D

etec

tor

LDPC SISO Decoder 1 1 -1 1

-1 -1 1 1

-1 -1 1 1

-1 1 1 3

10

00

11

September 2005

Douglas S. Chan et al., Cornell University

Slide 18

doc.: IEEE 802.11-05/0946r0

Submission

MUD 2: Iterative detection and decoding• Advantages:

– Each user’s code are the same• Simplifies multiaccess LDPC design

• A simpler implementation

– BP decoding of LDPC can be done in parallel

• Implementation issues:– Requires fixed length codewords– Codes designed only for a fixed number of users

• We can put limit to the number of users associating to AP

– AP needs to tell each user what interleaving scheme to use during association– Requires good codes to combat code expansion (CE)– No MPR when AP transmits

September 2005

Douglas S. Chan et al., Cornell University

Slide 19

doc.: IEEE 802.11-05/0946r0

Submission

Performance of MUD enhanced 802.11

September 2005

Douglas S. Chan et al., Cornell University

Slide 20

doc.: IEEE 802.11-05/0946r0

Submission

Simulation details

• Used 802.11a overheads in PHY preamble and MAC headers

• Asymptotic (or saturated) user traffic– Stress system to find asymptotic performances

– Emulates current user traffic scenarios

• Assume every user’s packet is corrupted if only one bit of a single user is corrupted– Worse case performance for MPR-enabled 802.11a

September 2005

Douglas S. Chan et al., Cornell University

Slide 21

doc.: IEEE 802.11-05/0946r0

Submission

Cross Layer Design• The MAC’s protocol is the same for each PHY layer

• While exponential backoff works, it’s not matched with the underlying PHY layer– [Calì] showed that the current MAC is not optimal for 802.11

• We show in the next slide the goodput for using the optimal transmission probability p

Backoffslots

Transmission

p p p p

September 2005

Douglas S. Chan et al., Cornell University

Slide 22

doc.: IEEE 802.11-05/0946r0

Submission

Performance of MUD enhanced 802.11 with Cross-layer (matched) MAC

September 2005

Douglas S. Chan et al., Cornell University

Slide 23

doc.: IEEE 802.11-05/0946r0

Submission

Conclusion

• A MPR-enabled 802.11 architecture is proposed

• Two MPR methods that preserve the underlying PHY layer are discussed

• Pointed out the importance of cross-layer design via a matched MAC

• Showed significant performance improvements for MPR enabled 802.11 and with matched MAC

September 2005

Douglas S. Chan et al., Cornell University

Slide 24

doc.: IEEE 802.11-05/0946r0

Submission

Acknowledgements

• Bruce Kraemer

• Teik-Kheong “TK” Tan

• Anuj Batra

September 2005

Douglas S. Chan et al., Cornell University

Slide 25

doc.: IEEE 802.11-05/0946r0

Submission

References• [Calì] F. Calì, M. Conti and E. Gregori, “Dynamic Tuning of the IEEE 802.11

Protocol to Achieve a Theoretical Throughput Limit,” IEEE/ACM Trans. Networking, vol. 8, pp. 785-799, 2000.

• [Chan1] D.S. Chan, T. Berger and L. Tong, “On the Stability and Optimal Decentralized Throughput of CSMA with Multipacket Reception Capability”, Allerton Conference on Communication, Control, and Computing, 2004.

• [Chan2] D.S. Chan and T. Berger, “Performance and Cross-Layer Design of CSMA for Wireless Networks with Multipacket Reception,” Proc. of Asilomar Conf. on Signals, Systems and Computers, 2004.

• [Ghez] S. Ghez, S. Verdú, and S. C. Schwartz, “Optimal decentralized control in the random-access multipacket channel,” IEEE Trans. Automat. Contr., 1989.

• [Palanki] R. Palanki, A. Khandekar and R. McEliece, “Graph based codes for synchronous multiple access channels,” Allerton Conference on Communication, Control, and Computing, 2001.

• [Sanderovich] A. Sanderovich, M. Peleg and S. Shamai, “LDPC Coded MIMO Multiple Access With Iterative Joint Decoding,” IEEE Trans. Info. Theory, 2005.